A Computer Science graduate with a passion for Deep-Learning,Blockchain and Linux.
I'm a recent CSE graduate from KUET with a strong foundation in programming and software development. My favourite topics include deep neural nets, blockchain technology, Database and Linux. I did my undergraduate thesis in NLP where I developed and trained my custom encoder-only model (DMoEBERT) for natural language inference (NLI) from scratch and outperformed BERT-base and RoBERTa-base in efficiency and resource-usage.
I am comfortable working with text, image or audio data and enjoy experimenting with models to make them faster and more efficient, I am also passionate about blockchain and tinkering with Linux.
I love to learn and love to teach. I share my knowledge thorough medium articles.
Implemented a Denoising Diffusion Probabilistic Model (DDPM) using PyTorch, Attention-enhanced U-Net for reverse-denoising, a Variational Autoencoder (VAE) for latent space compression, two-stage optimized training approach.
View ProjectA decentralized file sharing and storage platform capable of processing any file-format e.g., mp3, mp4, jpeg, py, zip..
View ProjectAn AI-Powered Course Recommendation System. Sentence Transformers for semantic embeddings and ChromaDB for efficient vector storage, integrated with a FastAPI backend and Streamlit UI.
View ProjectA project i am currently working on.An in-memory key-value store similar to redis.
View Project